case study
The 2024 → 2026 self-benchmark
drafted by the agent half · human review pending
10× / ⅒
roughly ten times the scope breadth, at roughly a tenth of the cash — the same person, six months, two years apart
basis: the founder’s own ledger; self-reported, not independently verifiable
How much does working with frontier AI actually change what one person can ship? If you’re weighing that question for yourself or your business, most answers you’ll find are vendor benchmarks. This one is different: the same person, the same calendar length, two years apart — which holds the hardest variable constant.
Everything on this page is the founder’s own ledger — not independently verifiable, and published anyway, framed exactly that way, because a controlled self-comparison is rarer and more honest than an industry number.
One thing this comparison is not: a case against teams. In 2024 the founder raised money, formed a team, and spent six months building a preference-intelligence product. One working MVP shipped — and so did an education. That team, and the years around it, are where the prerequisites were learned: environments, deployment, versioning, hosting, how real software actually gets out the door. The 2026 column exists because of the 2024 column, not despite it. What changed in between was the tools — and what the tools could do with an already-paid tuition.
the comparison
Self-reported, tuition included
| 2024 — first attempt, with a team | 2026 — same person, experience + frontier tools | |
|---|---|---|
| Time | 6 months from team formation | 6.5 months |
| Cash | ~$30K, plus heavy sweat equity from a team working at below-market rates | subscription-scale — low thousands |
| Working context | founder + team — with all the real coordination any group carries | one person + AI, context held in a persistent memory system |
| Output | one working MVP — the hard first one, and the education that came with it | a full real-estate data stack, a running client practice, a research corpus, always-on infrastructure, and creative work |
Roughly ten times the scope breadth at roughly a tenth of the cash. The compounding is the story: experience earned with people, multiplied by tools that arrived later.
take it with you
The tuition compounds
If you are where the 2024 column was — building the hard first thing, learning the prerequisites as you go — the tuition is worth paying, and it compounds more than it ever has: the tools that arrive tomorrow multiply whatever foundation you build today. And if you already carry years of hard-won experience, the 2026 column is a calibration for what it might now be worth.
what isn’t true yet
Read this in the same register as the claims
No third party has audited either column. The 2026 column’s breadth can be partially inspected through this site’s inventory and open numbers, but the comparison itself remains one person’s honest accounting of their own two attempts. Weigh it as exactly that.